1 Financial support for M.H.H. was provided by theUniversity of Montana Honors College in the form ofa Gordon and Anna Watkins Scholarship. This studywas conducted by M.H.H. in partial fulfillment of therequirements of the degree of Bachelor of Science ofWildlife Biology in the Division of Biological Sciences ofthe University of Montana. Manuscript acceptedI September 1997.

2 Division of Biological Sciences, University ofMontana, Missoula, Montana 59801.

3 Direct correspondence to Pacific Rim Research,1061 Kokomo Road, Ha'iku, Hawai'i 96708; E-mail:march@aloha.net.

PRISTINE HAWAllAN STREAMS host a nativeamphidromous fauna composed of endemiccrustaceans, gobies, and gastropods (includ­ing Neritina granosa Sowerby) (Ford andKinzie 1982, Kinzie and Ford 1982, Kinzie1988, Anonymous 1990). All of these speciesare considered "freshwater amphidromolls"(sensu McDowall 1992). Freshwater amphi­dromy is a subcategory of diadromy, wherebyindividuals live out the adult phase of theirlife cycle, including breeding, within a fresh­water stream. The larvae produced are carriedby the stream waters to the ocean where

Pacific Science (1998), vol. 52, no. 3: 237-249© 1998 by University of Hawai'i Press. All rights reserved

Population Genetics and Pattern of Larval Dispersal of the EndemicHawaiian Freshwater Amphidromous Gastropod Neritina granosa

(Prosobranchia: Neritidae) 1

MARC H. HODGES2 ,3 AND FRED W. ALLENDORF2

ABSTRACT: Protein electrophoresis was used to study the population geneticsof the endemic Hawaiian freshwater amphidromous gastropod Neritina granosaSowerby. The genetic information was used to infer the pattern and degreeof planktonic larval dispersal. Samples were taken from 12 streams locatedthroughout the Hawaiian Archipelago during July, August, and September1991. Overall mean heterozygosity was 0.052. Heterozygote deficiency wascomparable with that found in other mollusks and marine invertebrates. Geneflow was substantial and was generally sufficient to maintain similar allele fre­quencies among stream populations. An island model of migration was in­dicated. However, significant heterogeneity among populations was observedand was due primarily to three geographically disparate streams. Causes ofdeficiency and heterogeneity remain unknown. Demographic information sug­gests that, although high from a genetic point of view, the rate of migrationcalculated from gene flow might be insufficient to affect demographic processesin large populations of N. granosa.

they develop as part of the marine plankton.Settlement-age larvae leave the ocean andreturn to the stream habitat to developinto adults. Adults do not migrate betweenstreams. The only method of dispersal be­tween streams is via the larval phase. Thus,in the Hawaiian freshwater amphidromousfauna, each stream population is a discreteentity. The stream populations are linked bysome degree of larval dispersal to form ametapopulation.

This fauna is an important component ofHawai'i's biodiversity. It was used for sub­sistence in ancient times by Native Hawai­ians. Today, certain species within this faunacontinue to be valued by Native Hawaiiansand Hawai'i's rural residents as a source ofsubsistence food items and recreation. Un­fortunately, Hawai'i's native stream faunahas declined during historical times. Primarycauses are habitat degradation (Maciolek1975, 1978, Parrish et al. 1978, Ford andYuen 1988) and extensive invasion by alienspecies (Maciolek 1975, 1984, Kinzie andFord 1977, 1982, Timbol et al. 1980, Kinzie1988, 1992). Despite their value, compara-

237

238

tively little is known of the populationgenetics of these species. Practically nothingis known of the pattern and degree of dis­persal of their planktonic larvae.

Understanding the connectivity of popu­lations within a metapopulation is a keychallenge in aquatic science (Doherty 1991).Likewise, information on larval dispersalis integral to conservation planning forHawai'i's stream fauna (Ford and Yuen1988, Anonymous 1990, Fairweather 1991).Knowledge of both the pattern and degree ofdispersal is necessary for identification of thesources of recruits and the size and locationof management units (e.g., Dizon et al. 1992).Persistence of the metapopulation itself de­pends on dispersal (Gilpin 1987). Where sizeand numbers of populations have been re­duced, information on dispersal is critical topredicting persistence of individual popu­lations and the metapopulation. (Reviews ofmetapopulation theory are provided by Kar­eiva {1990] and Hanski and Gilpin [1991].)However, the degree to which the larvae ofHawai'i's native stream fauna return to theirnatal stream or recruit to nonnatal streams isunknown for any of the species.

Direct measurement of larval dispersalis notoriously difficult (Leis 1991, Grosbergand Levitan 1992). Genetic studies have beenused regularly with other fauna to infer thespatial pattern of dispersal among popula­tions and to calculate the number of migrantsper generation (Waples 1987).

Previous work on the population geneticsof the Hawaiian amphidromous fauna wasundertaken by Kinzie and Ford (1982)(N. granosa) and Fitzsimons et al. (1990)(the goby Lentipes conca/or). Both studiesconcluded that gene flow is sufficient toprevent genetic divergence among streams(= populations) throughout Hawai'i. How­ever, the number of populations sampled wassmall (concern about the status of L. conca/orprecluded an extensive survey). Neither in­vestigation described the pattern of gene flowamong populations and islands, nor assessedthe probable proportion of migrants in indi­vidual populations. This study was under­taken to further describe the population

PACIFIC SCIENCE, Volume 52, July 1998

genetics of N. granosa and to use this infor­mation to assess the spatial pattern and de­gree of larval dispersal among stream popu­lations occurring throughout the HawaiianArchipelago.

Locally known as hlhlwai, N. granosa isthe largest of the Hawaiian freshwaterneritids (Maciolek 1978). It is a coveted sub­sistence food item for Native Hawaiians.Most of its basic biology is known from Ford(1979). The largest individuals reach 50 mmin shell length. Sexual maturity occurs atabout 12 mm shell length. The species exhibitsseparate sexes and iteroparity. Individualsmay live from 1 to 7 yr. Sessile egg capsulesare affixed directly to the substrate. Afterhatching, larvae are swept to the sea. Theshells of the veliger larvae are 150-175 f-lmin diameter at hatching and 540-660 f-lm atsettlement. Both reproduction and recruit­ment occur year-round but with some peri­odicity (Ford 1979, Hodges 1992; M.H.H.,unpubl. obs.; C. Way, pers. comm.). Basedon growth studies in the laboratory, Ford(1979) estimated the length of time spent bythe larvae in the marine environment at closeto a year. However, the actual length of timeis unknown and seems likely to be less thanthat. Sometime after returning to a streamfrom the ocean, individuals ranging fromabout 5 to 15 mm shell length spreadthroughout the adult habitat by migratingupstream in long trains (Ford 1979, Hodges1992).

MATERIALS AND METHODS

M.H.H. collected samples of N. granosafrom three streams on each of four islands(Figure 1). These are HanakapI'ai (159° 36'W,22° 12'N), Wainiha (159° 32'W, 22° 12'N),and Hanalei (159° 30'W, 22° 13'N) onKaua'i; Waioho'okalo (156° 54' W, 21° 10' N),Wailau (156° 50'W, 21° 10'N), and Kawai­nui (156° 48'W, 21° IO'N) on Moloka'i;Honomanu (156° 10' W, 20° 52' N), Wai­ohue (156° 07' W, 20° 49' N), and HanawI(156° 06' W, 20° 49' N) on Maui; and Nanue(155° 09' W, 19° 55' N), Hakalau (155° 07' W,

Population genetics of Neritina granosa-HODGES AND ALLENDORF 239

1 2 3

~, Kaua';

tfl ON;';hau

Sampling Sites1. Hanakapi'ai2. Wainiha3. Hanalei4. Waioho'okalo5. Wailau6. Kawainui7. Honomanu8. Waiohue9. Hanawi10. Nanue11. Hakalau12. Kapue

O'ahu:;../\ 4 5 6

~ \// 789

Moloka'; ~S5\// .<"'"'-. MaUl

Lana'; U~ 101112

Kaho'olawe \11 20"

Hawa;';

FIGURE I. Location of sample streams.

19° 54'N), and Kapue (155°05'W, 19° 47'N)on Hawai'i. Each stream was sampled overthe span of a few hours. Sampling occurredbetween July and September 1991.

Samples were collected at the terminus ofeach stream except Hanalei, where sampleswere collected where the first riffle occursapproximately 5.5 Ian inland and abovethe range of tidal influence. Straight-line dis­tances among the termini of sample streamsrange from 2 to 540 Ian.

To examine within-stream heterogeneity,samples were collected approximately 1500 mand 700 m upstream of the termini in Hana­kapI'ai and HanawI Streams, respectively,for comparison with samples taken at thetermini of these two streams. The samplescollected at all locations exhibited the "lowerstream" shell morph of Ford (1979).

Collection of samples in streams wasrestricted to as small an area as possible

(usually ca. 70 m2). Only adults (greater than12 mm shell length) were collected. Sampleswere transported alive to freezers, thenbrought to the laboratory for analysis.

Protein Electrophoresis

We used both foot and hepatopancreastissue for assay at each presumptive locus.Tissues were homogenized in 10 ml of 0.25gfliter pyridoxal-5-phosphate grinding solu­tion and centrifuged at 5000 rpm for 5 min.Gels were made with 12% hydrolyzed Sigmastarch. Gel buffer systems used were AC (gelbuffer: 0.002 M citric acid, pH adjusted withN-(3-aminopropyl)-morpholine and citricacid; electrode buffer: 0.04 M citric acid, pH6.1 [Clayton and Tretiak 1972]); MF (stocksolution: 0.9 M tris, 0.5 M boric acid, 0.02 MEDTA, pH 8.7; gel buffer: 1: 20 dilution ofstock solution; electrode buffer: 1: 5 dilution

240 PACIFIC SCIENCE, Volume 52, July 1998

TABLE I

ENZYMES USED IN ELECTROPHORETIC SURVEY OF N. granosa IN HAWAI'I

ENZYME"

Aspartate aminotransferase (AAT)Acid phosphatase (ACP-I)

(ACP-2)Glyceraldehyde-3-phosphate dehydrogenase (G3PDH)Glucose-6-phosphate isomerase (GPI)Isocitrate dehydrogenase (IDHP-I)

(IDHP-2)Leucine aminopeptidase (LAP)Malate dehydrogenase (MDH)Mannose-6-phosphate isomerase (MPI)Phenylalanyl-proline (PEP)6-phosphogluconate dehydrogenase (PGDH)Phosphoglucomutase (pGM)Sorbitol dehydrogenase (SDH)Superoxide dismutase (SOD)

a Nomenclature from Shaklee et a~. (1990).• See text for buffer descriptions.

of stock solution [Market and Faulhaber1965)); RW (gel buffer: 0.03 M tris, 0.005 Mcitric acid, pH 8.5; electrode buffer: 0.06 Mlithium hydroxide, 0.03 M boric acid, pH 8.1[Ridgway et al. 1970)). We followed the elec­trophoretic methods of Leary and Booke(1990) and the histochemical staining tech­niques of Harris and Hopkinson (1976).

We assembled a list of 50 enzymes fromthe literature that had been successfullyscreened in other gastropods. Fifteen e.f thoseenzymes stained reliably, and we scored thesefor approximately 30 individuals per streampopulation. Enzymes and buffer systems aregiven in Table 1. Isozymes of multilocussystems are numbered from cathode toanode. Alleles are numbered according torelative electrophoretic mobility and pre­sented according to approximate frequencyof occurrence: AAT*100, *110, *80, *90;A CP-1 *-100, *-180, *-60; GPI*100, *225,*25; IDHP-2*100, *5, *200, *120; MPI*100,*115, *60; PEP*100, *109, *97; PGDH*100,*275, *-50; PGM*100, *108, *90, *95, *80;SDH*100, *86, *106.

We used the BIOSYS-l computer pro­gram (Swofford and Selander 1989) toanalyze electromorph frequencies. Levene's(1949) correction for small sample size was

E.C. BUFFER SYSTEM"

2.6.1.1 AC (pH 6.1)3.1.3.2 AC (pH 6.1)If If

1.2.1.12 AC (pH 6.9)5.3.1.9 RW1.1.1.42 AC (pH 6.9)If If

3.4.11.1 RW1.1.1.37 AC (pH 6.9)5.3.1.8 RW3.4.1.1 MF1.1.1.44 AC (pH 6.9)5.4.2.2 RW1.1.1.14 RW1.15.1.1 RW

applied in tests for deviation from Hardy­Weinberg proportions.

We quantified heterozygote deficiencywith D* = (Ho - He)/He, where Ho is ob­served heterozygosity and He is expectedheterozygosity. This is simply Wright'sfixation index with a negative sign, butemphasizes deficiency (Zouros et al. 1980).Homogeneity of heterozygote deficiencieswas tested by treating D* as a correlationcoefficient (Sokal and Rolf 1981, Gaffneyet al. 1990).

RESULTS

Of the 15 loci assayed in N. granosa, ninewere polymorphic and the overall meanheterozygosity was 0.052 (Table 2). Eightof the 12 populations deviated significantlyfrom Hardy-Weinberg proportions at someloci (Table 2). These deviations occurred atsix of nine polymorphic loci. Deviations wereobserved in 17 out of 83 tests, more thanexpected by chance alone at P < .05. Allsignificant deviations were as deficiencies ofheterozygotes. The mean value of D* was-0.122.

There was no detectable difference in fre-

TABLE 2

ALLELE FREQUENCIES AT NINE LOCI IN 12 POPULATIONS OF N. granosa IN HAWAI'I(HI' Is OBSERVED HETEROZYGOSITY, He Is THE HETEROZYGOSITY EXPECTED UNDER HARDy-WEINBERG EQUILIBRIUM,

AND D* Is HETEROZYGOTE DEFICIENCY)

LOCUS POPULATION n ALLELE FREQUENCY HI' H,. D*

AAT *100 *1l0 *80 *90

HanakapI'ai 30 0.733 0.233 0.033 0.133 0.407 -0.672***Wainiha 28 0.679 0.304 0.018 0.286 0.447 -0.361Hanalei 30 0.550 0.450 0.433 0.495 -0.125Waioho'okalo 30 0.767 0.217 0.017 0.300 0.365 -0.178Wailau 30 0.733 0.217 0.033 0.017 0.267 0.414 -0.356Kawainui 24 0.458 0.458 0.042 0.042 0.625 0.576 0.084Honomanu 30 0.900 0.050 0.050 0.200 0.185 0.081Waiohue 30 0.700 0.283 0.017 0.267 0.429 -0.379*HanawI 30 0.667 0.317 0.017 0.267 0.455 -0.414*Nanue 28 0.672 0.293 0.034 0.379 0.461 -0.177Hakalau 24 0.688 0.250 0.042 0.021 0.292 0.463 -0.370Kapue 30 0.667 0.333 0.333 0.444 -0.250

mean = -0.260

ACP-l *-/00 *-/80 *-60

HanakapI'ai 29 0.983 0.017 0.034 0.034 0.018Wainiha 29 0.948 0.034 0.017 0.103 0.099 0.042Hanalei 29 0.966 0.034 0.067 -1.00***Waioho'okalo 30 0.983 0.017 0.033 0.033 0.017Wailau 30 1.00!(.awainui 26 0.885 0.096 0.019 0.231 0.208 0.110Honomanu 30 1.00Waiohue 30 0.983 0.017 0.033 0.033 0.017HanawI 30 0.950 0.050 0.100 0.095 0.053Nanue 30 1.00Hakalau 30 1.00Kapue 30 1.00

mean = -0.106

GPI *100 *225 *25

HanakapI'ai 30 0.983 0.017 0.033 0.033 0.017Wainiha 30 1.00Hanalei 30 0.967 0.033 0.064 -1.00***Waioho'okalo 30 0.967 0.017 0.017 0.067 0.065 0.026Wailau 30 0.950 0.050 0.100 0.095 0.053Kawainui 30 0.983 0.017 0.033 0.033 0.017Honomanu 30 0.983 0.017 0.033 0.033 0.017Waiohue 30 0.983 0.017 0.033 0.033 0.017HanawI 30 0.967 0.033 0.067 0.064 0.034Nanue 30 0.917 0.067 0.017 0.167 0.155 0.075Hakalau 30 0.983 0.017 0.033 0.033 0.017Kapue 30 0.967 0.017 0.017 0.067 0.065 0.026

mean = -0.064

IDHP-2 *100 *5 *200 *120

HanakapI'ai 30 0.950 0.017 0.017 0.017 0.100 0.097 0.034Wainiha 29 0.983 0.017 0.034 0.034 0.018Hanalei 27 0.963 0.037 0.074 0.071 0.038Waioho'okalo 30 1.00Wailau 30 1.00Kawainui 30 0.967 0.017 0.017 0.067 0.065 0.026Honomanu 30 0.983 0.017 0.033 0.033 0.017Waiohue 30 1.00

TABLE 2 (continued)

LOCUS POPULATION n ALLELE FREQUENCY Ho He D*

HanawI 29 0.966 0.034 0.069 0.067 0.036Nanue 30 1.00Hakalau 30 0.967 0.017 0.017 0.067 0.065 0.026Kapue 30 0.983 0.017 0.033 0.033 0.017

mean = 0.027

MPI *100 *115 *60

HanakapI'ai 30 0.983 0.017 0.033 0.033 0.017Wainiha 29 0.931 0.052 0.017 0.069 0.130 -0.470*Hanalei 27 0.889 0.111 0.074 0.198 -0.625**Waioho'okalo 28 0.857 0.071 0.071 0.214 0.255 -0.160Wailau 26 0.904 0.058 0.038 0.077 0.178 -0.568***Kawainui 29 0.929 0.036 0.036 0.071 0.135 -0.472Honomanii 30 0.967 0.033 0.067 0.064 0.034Waiohue 29 0.914 0.034 0.052 0.172 0.161 0.Q70HanawI 30 0.933 0.050 0.017 0.067 0.126 -0.471*Nanue 30 0.950 0.050 0.100 0.095 0.053Hakalau 29 0.948 0.017 0.034 0.103 0.099 0.042Kapue 30 0.850 0.100 0.050 0.233 0.265 -0.119

mean = -0.222

PEP *100 *109 *97

HanakapI'ai 30 0.867 0.083 0.050 0.267 0.239 0.114Wainiha 29 0.948 0.034 0.017 0.034 0.099 -0.653***Hanalei 27 0.950 0.050 0.100 0.095 0.053Waioho'okalo 28 0.983 O.oI7 0.033 0.033 0.017Wailau 26 0.875 0.125 0.107 0.219 -0.510**Kawainui 29 0.900 0.100 0.067 0.180 -0.630**Honomanii 30 0.950 0.017 0.033 0.100 0.096 0.040Waiohue 29 0.950 0.017 0.033 0.100 0.096 0.040HanawI 30 0.942 0.038 0.019 0.115 0.110 0.047Nanue 30 0.950 0.050 0.100 0.095 0.053Hakalau 29 0.883 0.083 0.033 0.233 0.212 0.102Kapue 30 0.933 0.033 0.033 0.067 0.127 -0.474'

mean = -0.150

PGDH *100 *275 *-50

HanakapI'ai 27 1.00Wainiha 30 1.00Hanalei 27 1.00Waioho'okalo 30 0.967 0.017 0.017 0.067 0.065 0.026Wailau 30 0.983 0.017 0.033 0.033 0.017Kawainui 30 0.950 0.033 0.017 0.033 0.096 -0.653***Honomanii 30 0.983 O.oI7 0.033 0.033 O.oI7Waiohue 30 0.983 O.oI7 0.033 0.033 0.017HanawI 30 0.933 0.067 0.067 0.124 -0.464*Nanue 30 0.983 0.017 0.033 0.033 0.017Hakalau 30 1.00Kapue 30 0.933 0.017 0.050 0.067 0.126 -0.471 *

mean = -0.187

PGM *100 *108 *90 *95 *80

HanakapI'ai 30 0.967 0.017 0.017 0.067 0.065 0.026Wainiha 30 0.967 0.033 0.067 0.064 0.034Hanalei 30 0.950 0.017 0.017 0.017 0.100 0.097 0.034Waioho'okalo 30 1.00

Population genetics of Neritina granosa-HODGES AND ALLENDORF 243

TABLE 2 (continued)

LOCUS POPULATION n ALLELE FREQUENCY Ho He D*

Wailau 30 0.950 0.017 0.017 0.017 0.100 0.097 0.034Kawainui 30 0.983 0.017 0.033 0.033 0.017Honomanii 30 0.983 0.017 0.033 0.033 0.017Waiohue 30 0.917 0.067 0.017 0.167 0.155 0.075Hanawi 30 0.983 0.017 0.033 0.033 0.017Nanue 30 0.967 0.033 0.067 0.064 0.034Hakalau 30 0.983 0.017 0.033 0.033 0.017Kapue 30 0.967 0.017 0.017 0.067 0.065 0.026

mean = 0.030

SDH *100 *86 *106

Hanakapi'ai 23 1.00Wainiha 24 0.958 0.021 0.021 0.083 0.081 0.032Hanalei 30 1.00Waioho'okalo 30 1.00Wailau 29 0.983 0.017 0.034 0.034 0.018Kawainui 30 1.00Honomanii 30 1.00Waiohue 29 1.00Hanawi 30 1.00Nanue 29 1.00Hakalau 30 1.00Kapue 30 1.00

mean = 0.025

Chi-square test for deviation from Hardy-Weinberg expectations: *, P < .05; **, P < .01; ***, P < .001.

TABLE 3

SUMMARY OF F-STATISTICS FOR 12 POPULATIONS OFN granosa IN HAWAI'I

NO. OFLOCUS ALLELES F(IS) F(ST) F(lT)

AAT 4 0.264 0.049* 0.300Aep-I 3 0.058 0.038* 0.094OPI 3 0.059 0.016 0.074IDHP-2 4 -0.029 0.015 -0.014MPI 3 0.263 0.020 0.278PEP 3 0.173 0.021 0.190PODH 3 0.324 0.024* 0.341POM 5 -0.038 0.016 -0.022SDH 3 -0.028 0.023 -0.004Mean 0.198 0.034* 0.225

* Fst significantly >0: chi-square test (Workman andNiswander 1970).

quencies between upstream and downstreamsamples (Hanakapi'ai, df= 18, P= .248,X2 =21.6; Hanawi, df=14, P=.619, X2 =11.8). This suggests genetic homogeneity

within streams among individuals of thelower shell morpho

Wright's (1951) F-statistics, which providean index of the degree of genetic subdivision,are given in Table 3. Fst at AAT*, ACP-l*,and PGDH* is significantly greater than zero(P < .05, chi-square test [Workman and Nis­wander 1970]). Average Fst (0.034) was alsosignificant. This indicates significant geneticheterogeneity among the 12 stream popula­tions. Pairwise chi-square tests among allpopulations resulted in 12 differences outof 66 comparisons, more than expected bychance at P < .05. All differences involvedHanalei (Kaua'i), Kawainui (Moloka'i), andHonomanii (Maui). Thus the significance ofoverall Fst is largely due to differencesamong the samples from Hanalei, Kawainui,and Honomanii Streams, and between theseand the remaining populations.

The allele frequencies that cause heteroge­neity are clear at AAT* and ACP-l*. The

244 PACIFIC SCIENCE, Volume 52, July 1998

eKaua'j o Moloka'j o Maui fj. Hawai'j

Honomanuo

1.3 1.35 1.41.25

e

e 0o fj.

oeHanalei

Kawainuio=

.28 J~.L-~L...-....o---"'----'-~---'-~""""""'---'---'--~....L...-_-'----+

.26

.24

.22

Cf) .2

~ .18

.16

.14

.12

.1

.08 t;--Y--;;-~;;-~;-~~.----,--,,.........,r--r~OL...,----y-....,.-....,.-....,.-....,.--.--!.85 .9 .95 1 1.05 1.1 1.15 1.2

pes I

. FIGURE 2. Plot of first two pri~cipal component scores of allele frequencies at nine polymorphic loci in 12 popu­latIOns. Different symbols denote Islands on whIch populations occur. The three populations that are genetically dis­tmct are labeled.

allele indicated by AAT*110 is at high fre­quency in Hanalei and Kawainui and is rare!n Honomanu (Table 2). That by ACP*-180IS ~t ~omparatively high frequency in Ka­walllUi. Such notable differences are not ap­parent at PGDH*.

A principal components analysis per­formed on the covariance of the frequency ofthe c~mmon alleles indicated a lack of geo­graphic s.tructure. The first principal compo­nent, which accounts for 69.8% of the vari­ance in allele frequencies, is plotted againstthe second principal component, which ac­counts for an additional 13.4% (Figure 2).The first component (PC 1) is based primar­ily on differences at AAT* and ACP-l*. Thesecond (pC 2) is based on differences atMP/* and ~E~*. The first component sepa­rates KawalllUi and Honomanu, and to alesser extent Hanalei from the central cluster.Despite this, the genetic structure is consis­tent with an island model of migration.

DISCUSSION

Genetic Variation within Populations

The heterozygote deficiency observed inN. granosa is common in mollusks (Volckaertand Zouros 1989) and marine invertebrates(Grant and Lang 1991). Gartner-Kepkay etal. (1980, 1983) reported a mean D* of-0.492 and -0.396 in the mussel Mytilusedulis. Data from Zouros et al. (1980) gave amean D* of -0.298 for oysters (Crassostreavirginica). The scallop Placopecten magel­lanicus exhibited much lower D* (-0.065[Volckaert and Zouros 1989]). At -0.122

D* . 'mean III N. granosa was intermediate inrelation to those other species.

Heterozygote deficiency in marine in­vertebrates is a contentious issue. Hypothesesmost often presented are the Wahlund effect(genetic subdivision within the sample), nullalleles (alleles that fail to appear in the elec­trophoresis process), nonrandom mating, andnatural selection (Gartner-Kepkay et al. 1980,1983, Zouros et al. 1980, 1988, Mallet et al.

Population genetics of Neritina granosa-HODGES AND ALLENDORF 245

.05~------------------------..,

•• •o .

-.1

D*

•-.2 •

••

-.3.01 .02 .03 .04 .05

Fst

FIGURE 3. Mean heterozygote deficiency plotted against Fst at each locus (simple regression, one-tailed test:1'2 = 0.345, P < .05).

1985, Foltz 1986, Volckaert and Zouros 1989,Gaffney et al. 1990).

Though stream habitats are discrete,genetically distinct cohorts might occurwithin the same stream (e.g., Johnson andBlack 1984, Mallet et al. 1985), thus causinggenetic subdivision within the stream sampleand concomittant heterozygote deficiency(the Wahlund effect). If a WaWund effectis caused by distinct larval cohorts withinsamples, deficiency should be stronger atthose loci that show more differentiationamong populations (Volckaert and Zouros1989). A plot of D* for each locus averagedamong populations against Fst at that locussuggests a negative trend (Figure 3). Thissupports the possibility of a Wahlund effect.Note however that this trend is weak and isheavily influenced by the single data pointcorresponding to AA T*.

If null homozygotes are lethal or are ex­cluded from the sample the expected fre­quency of null alleles in N. granosa givenmean D* is 0.334 (Gaffney et al. 1990). Ifnulls are selectively neutral and only one

null allele segregates per locus, the frequencynecessary to account for the observed defi­ciency is of the same magnitude. Theseestimates are orders of magnitude largerthan frequencies determined for Drosophila(Langley et al. 1981) and pine trees (Allen­dorf et al. 1982), but similar to frequenciesinferred by Foltz (1986) for an oyster. How­ever, it is unlikely that null alleles will besegregating at such high frequency at six ofnine loci.

Nonrandom mating will cause uniformdeficiency among loci (Lewontin andKrakauer 1973, Zouros et al. 1980). D* isheterogeneous at five of the 12 populations(HanakapI'ai, Wainiha, Hanalei, Wailau,Kawainui; tested as correlation coefficients,P < .05; see Materials and Methods). In ad­dition, although adults are very sedentaryand fertilization is internal, the observationof within-stream genetic homogeneity of theshell morph used in this assay, coupled withthe fact that recruits migrate throughoutthe adult habitat, suggest that nonrandommating is an unlikely cause of the deficiency.

246

Selection is difficult to document, espe­cially in retrospect. Mallet et al. (1985) foundevidence for selective larval mortality inmussels, but noted that the WaWund effect(e.g., Burton 1983) and inbreeding were mostoften invoked as causes of deficiency. BothMallet et al. (1985) and Zouros et al. (1988)stated that there is no consensus on the causeof heterozygote deficiencies. Similarly, it isnot clear what causes deficiency in N. granosa.

Among-Population Heterogeneity

Heterogeneity in a high gene flow systemsuggests selection. Gartner-Kepkay et al.(1980, 1983) observed heterogeneity inmussels despite the high gene flow expectedof planktonic larvae and 'suggested selectiveenvironmental effects. Campton et al. (1992)documented temporal and spatial geneticpatchiness in the conch Strombus gigas.Heterogeneity exists on temporal and micro­geographic scales in the limpet Siphonariajeanae (Johnson and Black 1984). Thoseauthors concluded that differentiation wascaused by unique histories of localized re­cruitment that may have been driven byselective processes operating on the larvae.

Either stream-specific selection acting onpostsettlement individuals within the streamhabitat or selection acting on larvae at seaand thereby producing genetically distinctrecruitment events could cause the heteroge­neity observed in N. granosa. If stream-spe­cific selection is the cause, the pattern of het­erogeneity will remain constant throughtime. On the other hand, if at-sea selection isthe cause, the locations of genetically distinctpopulations are likely to change. Time-seriesallele frequency data are needed to investi­gate the relative importance of stream­specific selection.

Genetic subdivision within populationscan cause heterogeneity between populations.Such subdivision would likely manifest itselfas heterozygote deficiency within populations(i.e., the Wahlund effect). Heterozygote defi­ciencies do occur in the samples, and theWahlund effect has not been eliminated as acause of such deficiency. However, there wasno detectable difference among populations

PACIFIC SCIENCE, Volume 52, July 1998

in the number ofloci that exhibited significantheterozygote deficiencies (chi-square = 8.715,df= II, P = .65; note that sparse cells requireextreme caution in interpreting this test).And, although significant deficiencies occurat Hanalei and Kawainui, D* is positive forall loci in Honomanii. Thus, it appears thatany genetic subdivision occurring withinthe populations is probably not a cause ofthe observed heterogeneity among thepopulations.

Neutral processes can cause significantheterogeneity among populations under con­ditions of considerably higher gene flow thanthat calculated for N. granosa (Allendorf andPhelps 1981). If drift is the cause of hetero­geneity, one would expect to see equal diver­gence among polymorphic loci. However, al­though the values of Fst are similar amongthe nine ploymorphic loci, Fst is significant atthree of them (Table 3).

In sum, there are a number of processesthat could contribute to the genetic hetero­geneity observed among the populations ofN. granosa. These include stream-specific se­lection acting on postsettlement individuals,at-sea selection acting on larvae and produc­ing genetically distinct recruitment events,and neutral processes. However, the relativeimportance of each cannot be determinedwith existing information.

The genetic structure (Figure 2) is consis­tent with an island model of migration. Thusmigrants originate from any stream through­out the Islands with equal probability.

Amount of Dispersal

A simple estimate of dispersal amongpopulations can be made using Wright's(1969) equation:

Fst = 1/(4Nem + I) (I)

and solving for Nem (Waples 1987, Slatkinand Barton 1989), where N e is the geneticallyeffective subpopulation size and m is theproportion of the subpopulation that is madeup of migrants. Dispersal is in terms of theabsolute number of migrants per generationNem (gene flow) and is independent of sub­population size (Allendorf and Phelps 1981).

Population genetics of Neritina granosa-HODGES AND ALLENDORF 247

Equation 1 assumes selective neutrality andan island model of migration. Thus, determi­nation of Nem from Fst rests on the assump­tion of selective neutrality. However, Nem isfairly robust to selection where the level ofgene flow is on the order of that determinedhere (Slatkin and Barton 1989). Likewise,Figure 2 demonstrates that the pattern ofdifferences in allele frequencies is consistentwith the island model. Overall Fst fromTable 3 is 0.034. Equation 1 yields 7.103 forNem. Approximately seven migrants aremoving among streams each generation.

Genetic Versus Demographic Importance ofMigration

From a genetic point of view the Nem(= 7) calculated for N. granosa is high andindicates sufficient gene flow to maintainsimilar allele frequencies (Hartl and Clark1989: 318). A little migration goes a long waytoward affecting genetic processes, and theamount of migration necessary to maintaingenetic homogeneity is independent ofsubpopulation size (Allendorf and Phelps1981). However, where populations are large,a correspondingly large amount of migrationis required to control, or even affect, demo­graphic dynamics.

Hodges (1992) reported populations of N.granosa ranging from 70,000 to 350,000 in­dividuals in three streams on Maui. Manyother stream populations are probably of asimilar size, and some may well be larger.The rate of migration (seven individuals pergeneration) inferred from these genetic datamight not be sufficient to affect demographicprocesses in such large populations of N.granosa.

ACKNOWLEDGMENTS

Sample collection required the assistanceof A. Shepard, B. Anderson, C. Kim, M.Reber, S. Liva, N. E. Aluli, H. Nala'ielua, L.and C. Costales, P. Thompson, J. Kainoa, A.Sageet, E. Misaki, C. Baldwin, F. Trusdell,D. Heacock, I. Hodges, J. Reber, J. Loui, J.Waxman, and P. Neiss. K. Knudsen, G.

Sage, R. Leary, S. Forbes, and G. Luikartassisted in the laboratory. C. Way, R. Kinzie,S. Hau, R. Nishimoto, and A. Yuen sharedinformation. V. Yap provided inspiration.Valuable comments were made by A. Sheldon,B. Holthuis, M. DybdaW, and two anony­mous reVIewers.

LITERATURE CITED

ALLENDORF, F. W., and S. R. PHELPS. 1981.Use of allelic frequencies to describepopulation structure. Can. J. Fish. Aquat.Sci. 38: 1507-1514.

ALLENDORF, F. W., K. L. KNUDSEN, andG. M. BLAKE. 1982. Frequencies of nullalleles at enzyme loci in natural popula­tions of ponderosa and red pine. Genetics100:497-504.

ANONYMOUS. 1990. Hawaii stream assess­ment: A preliminary appraisal of Hawaii'sstream resources. Hawai'i CooperativePark Services Unit, Western Region Nat­ural Resources and Research Division,Honolulu.

BURTON, R. S. 1983. Protein polymorphismsand genetic differentiation of marine in­vertebrate populations. Mar. BioI. Lett.4: 193-206.

CAMPTON, D. E., C. J. BERG, L. M. ROBINSON,and R. A. GLAZER. 1992. Genetic patchi­ness among populations of queen conchStrombus gigas in the Florida Keys andBimini. Fish. Bull. 90: 250-259.

CLAYTON, J. W., and D. R. TRETIAK. 1972.Amine citrate buffers for pH control instarch gel electrophoresis. J. Fish. Res.Board Can. 29: 1169-1172.

DIZON, A. E., C. LOCKYER, W. F. PERRIN,D. P. DEMASTER, and J. SISSON. 1992. Re­thinking the stock concept: A phylogeo­graphic approach. Conserv. BioI. 6: 24-36.

DOHERTY, P. J. 1991. Spatial and temporalpatterns in recruitment. Pages 261-293 inP. F. Sale, ed. The ecology of fishes oncoral reefs. Academic Press, San Diego.

FAIRWEATHER, P. G. 1991. Implications of'supply side' ecology for environmentalassessment and management. TrendsEcol. Evol. 6: 61-64.

248

FITZSIMONS, J. M., R. M. ZINK, and R. T.NISHIMOTO. 1990. Genetic variation in theHawaiian stream goby Lentipes concolor.Biochem. Syst. Ecol. 18: 81-83.

FOLTZ, D. W. 1986. Null alleles as a possiblecause of heterozygote deficiencies in theoyster Crassostrea virginica and otherbivalves. Evolution 40: 869-870.

FORD, J. I. 1979. Biology of a Hawaiian flu­vial gastropod Neritina granosa Sowerby(Prosobranchia: Neritidae). M.S. thesis,University ofHawai'i at Manoa, Honolulu.

FORD, J. I., and R. A KINZIE III. 1982. Lifecrawls upstream. Nat. Hist. 91: 60-67.

FORD, J. I., and A YUEN. 1988. Naturalhistory of Pelekunu Stream and its tribu­taries, island of Moloka'i, Hawai'i. Part I:Summary report. Nature Conservancy ofHawai'i, Honolulu.

GAFFNEY, P. M., T. M. SCOTT, R. K. KOEHN,and W. J. DIEHL. 1990. Interrelationshipsof heterozygosity, growth rate and hetero­zygote deficiency in the coot clam Mulinialateralis. Genetics 124: 687-699.

GARTNER-KEPKAY, K. E., L. M. DICKIE,K. R. FREEMAN, and E. ZOUROS. 1980.Genetic differences and environments ofmussel populations in the Maritime Prov­inces. Can. J. Fish. Aquat. Sci. 37: 775­782.

GARTNER-KEPKAY, K. E., E. ZoUROS, L. M.DICKIE, and K. R. FREEMAN. 1983. Ge­netic differentiation in the face of geneflow: A study of mussel populations froma single Nova Scotian embayment. Can. J.Fish. Aquat. Sci. 40:443-451.

GILPIN, M. E. 1987. Spatial structure andpopulation vulnerability. Pages 125-139in M. E. Soule, ed. Viable populations forconservation. Cambridge University Press,Cambridge.

GRANT, W. S., and M. LANG. 1991. Modeof larval development and genetic popu­lation structure in Nodilittorina africanaknysaensis (Prosobranchia: Littorinidae).Mar. BioI. (BerI.) 109: 479-483.

GROSBERG, R. K., and D. R. LEVITAN. 1992.For adults only? Supply-side ecology andthe history of larval biology. Trends EcoI.EvoI. 7: 130-133.

HANSKI, I., and M. GILPIN. 1991.

PACIFIC SCIENCE, Volume 52, July 1998

Metapopulation dynamics: Brief historyand conceptual domain. BioI. J. Linn. Soc.42:3-16.

HARRIs, H., and D. A HOPKINSON. 1976.Handbook of enzyme electrophoresis inhuman genetics. North-Holland Publish­ing, Amsterdam.

HARTL, D. L., and A G. CLARK. 1989. Prin­ciples of population genetics, 2nd ed.SunderlandfSinauer Associates, Boston.

HODGES, M. H. 1992. Population biologyand genetics of the endemic Hawaiianstream gastropod Neritina granosa (Pro­sobranchia: Neritidae): Implications forconservation. M.S. thesis, University of

. Montana, Missoula.JOHNSON, M. S., and R. BLACK. 1984. Pattern

beneath the chaos: The effect of recruit­ment on genetic patchiness in an intertidallimpet. Evolution 38: 1371-1383.

KAREIVA, P. 1990. Population dynamics inspatially complex environments: Theoryand data. Philos. Trans. R. Soc. Lond. B330: 175-190.

KINZIE, R. A, III. 1988. Habitat utilizationby Hawaiian stream fishes with referenceto community structure in oceanic islandstreams. Environ. BioI. Fishes 22: 179­192.

---. 1992. Predation by the introducedcarnivorous snail Euglandia rosea (Ferus­sac) on endemic aquatic lymnaeid snailsin Hawaii. BioI. Conserv. 60: 149-155.

KINZIE, R. A, III, and J. I. FORD. 1977. Alimnological survey of lower Palikea andPipiwai Streams, Kipahulu, Maui. Coop.Natl. Park Resour. Stud. Unit, Univ. Ha­waii, Tech. Rep. 17.

---. 1982. Population biology in smallHawaiian streams. Hawaii Coop. Fish.Res. Unit, Water Resour. Res. Cent.,Univ. Hawaii, Tech. Rep. 147.

LANGLEY, C. H., R. A VOELKER, A. J. LEIGHBROWN, S. OHNISHI, B. DICKSON, and E.MONTGOMERY. 1981. Null allele fre­quencies at allozyme loci in naturalpopulations of Drosophila melanogaster.Genetics 99: 151-156.

LEARY, R. F., and H. E. BooKE. 1990. Starchgel electrophoresis and species distinc­tions. Pages 141-170 in C. B. Schreck and

Population genetics of Neritina granosa-HODGES AND ALLENDORF 249

Lii

P. B. Moyle, eds. Methods for fish biology.American Fisheries Society, Bethesda,Maryland.

LEIS, J. M. 1991. The pelagic stage of reeffishes: The larval biology of coral reeffishes. Pages 183-227 in P. F. Sale, ed.The ecology of fishes on coral reefs. Aca­demic Press, San Diego.

LEVENE, H. 1949. On a matching problemarising in genetics. Ann. Math. Stat.20: 91-94.

LEWONTIN, R. c., and J. KRAKAUER. 1973.Distribution of gene frequency as a test ofthe theory of the selective neutrality ofpolymorphisms. Genetics 74: 175-195.

MACIOLEK, J. A. 1975. Limnological ecosys­tems and Hawaii's preservational plan­ning. Verh. Int. Ver. Limnoi. 19: 1461­1467.

---. 1978. Shell character and habitatof nonmarine Hawaiian neritid snails.Micronesica 14: 209-214.

--_-. 1984. Exotic fishes in Hawaii andother islands of Oceania. Pages 131-161 inW. R. Courtenay, Jr., and J. R. Stauffer,Jr., eds. Distribution, biology, and man­agement of exotic fishes. Johns HopkinsUniversity Press, Baltimore.

MALLET, A. L., E. ZOUROS, K. E. GARTNER­KEPKAY, K. R. FREEMAN, and L. M.DICKIE. 1985. Larval viability and hetero­zygote deficiency in populations of marinebivalves: Evidence from pair matings ofmussels. Mar. BioI. (BerI.) 87: 165-172.

MARKET, C. L., and I. FAULHABER. 1965.Lactate dehydrogenase isozyme patternsoffish. J. Exp. Zool. 159:319-335.

McDoWALL, R. M. 1992. Diadromy: Originsand definitions of terminology. Copeia (1):248-251.

PARRISH, J. D., J. A. MACIOLEK, A. S. TIMBOL,C. B. HATHAWAY, JR., and S. E. NORTON.1978. Stream channel modification inHawaii. Part D: Summary report FWS/OBS 78/19. U.S. Fish and WildlifeService, Honolulu.

RIDGWAY, G. J., S. W. SHERBURNE, and R.D. LEWIS. 1970. Polymorphisms in theesterases of Atlantic herring. Trans. Am.Fish. Soc. 99: 147-151.

SHAKLEE, J. B., F. W. ALLENDORF, D. C.

MORIZOT, and G. S. WHITT. 1990. Genenomenclature for protein-coding loci infish. Trans. Am. Fish. Soc. 119: 2-15.

SLATKIN, M., and N. H. BARTON. 1989. Acomparison of three indirect methods forestimating average levels of gene flow.Evolution 43: 1349-1368.

SOKAL, R. R., and F. J. ROHLF. 1981. Biom­etry. W. H. Freeman, San Francisco.

SWOFFORD,D. L., and R. B. SELANDER. 1989.BIOSYS-1: A computer program for theanalysis of allelic variation in populationgenetics and biochemical systematics. Re­lease 1.7. Illinois Natural History Survey,Chicago.

TIMBOL, A. S., A. J. SUTTER, and J. D.PARRISH. 1980. Distribution, relative abun­dance, and stream environment of Len­tipes concolor (Gill 1860), and associatedfauna in Hawaiian streams. Hawaii Coop.Fish. Res. Unit, Water Resour. Res.Cent., Univ. Hawaii, Coop. Rep. 5.

VOLCKAERT, F., and E. ZOUROS. 1989. Allo­zyme and physiological variation in thescallop Placopecten magellanicus and ageneral model for the effects of hetero­zygosity on fitness in marine molluscs.Mar. BioI. (Beri.) 103: 51-61.

WAPLES, R. S. 1987. A multispecies approachto the analysis of gene flow in marineshorefishes. Evolution 41 :385-400.

WORKMAN, P. L., and J. D. NISWANDER.1970. Population studies on southwesternIndian tribes. II. Local genetic differ­entiation in the Papago. Am. J. Hum.Genet. 22: 24-29.

WRIGHT, S. 1951. The genetical structure ofpopulations. Ann. Eugen. 15: 323-354.

---. 1969. Evolution and the genetics ofpopulations, Vol. II. The analysis of genefrequencies. University of Chicago Press,Chicago.

ZOUROS, E., M. ROMERO-DoLEY, and A. L.MALLET. 1988. Heterozygosity and growthin marine bivalves: Further data and pos­sible explanations. Evolution 42: 1332­1341.

ZOUROS, E., S. M. SINGH, and H. E. MILES.1980. Growth rates in osyters: An over­dominant phenotype and its possible ex­planations. Evolution 34: 856-867.